Oct., 1947
HEATCAPACITIES OF 1~-HEXANE AND
Benzamides, wepared in the customary manner from "
ally from ethanol hone (see Table 11). When equal weights of the pyrrolidine and phenyl isothiocyanate were mixed, considerable heat was evolved and the tiiixture usually solidified upon cooling. After being washed with 50% ethanol the phenylthioureas were crystallized from chloroform-ethanol. These derivatives are likrwize listed in Table 11.
[CONTRIBUTION No. 7 FROM
THE
z,%DIMETHYLBUTANE
2275
Summarv to a,p-unsaturated ketones. 3. Substituted pyrrolidines were the principal products when a number of y-nitroketones were reduced Over Raney both in the presence and absence of ammonia. GREENCASTLE, INDIANA
RECEIVED MARCH3, 1947
THERMODYNAMICS LABORATORY OF THE PETROLEUM EXPERIMENT STATION, BUREAUOF MINES, BARTLESVILLE, OKLA.]
Experimental Vapor Heat Capacities and Heats of Vaporization of n-Hexane and 2,Z-Dimethylbutane BY GUYWADDINGTON AND DONALD R. DOUSLIN The current thermodynamic program of the Petroleum and Natural Gas Division of the Bureau of Mines includes the measurement of the heat capacities of hydrocarbon vapors. Incidental to the determination of vapor heat capacities accurate values of the heats of vaporization are obtained. A recent publication2 describes the %ow calorimeter and cycling vaporizer used for determining the above properties. This paper gives the results of measurements of the vapor heat capacities and heats of vaporization of n-hexane and 2,2-dimethylbutane. Experimental The Apparatus and Experimental Method.The flow calorimeter and vapor cycling system have been described recently in considerable detail, hence only a brief account of the method and apparatus will be given here. A measured constant flow of hydrocarbon vapor is produced in a glass vaporizing vessel thermally isolated from its surroundings. Normally the vapor is passed through the calorimeter, then condensed and returned to the vaporizer as liquid. The boiling temperature is kept constant by carefully controlling the pressure of helium gas in the static portion of the system. The dependence of rate of flow on the power input to the vaporizer heater is determined by diverting the stream of vapor into a suitable collecting vessel for a measured time int.erva1 and weighing the condensed sample of hydrocarbon. From properly corrected values of mass, power and time, accurate values of the heat of vaporization are obtained. To ascertain the heat capacity of the vapor the temperature rise produced in the flowing vapor by the measured power input to the calorimeter heater is determined. Two platinum resistance thermometers situated directly in the vapor . (1) Approved for publication by the Director of the Bureau of &lines, U S Department of the Interior. Not copyrighted. ('2) Guy Waddington, Samuel S. Todd and Hugh M. Huffman, THISJ D U I W A L , 69, 22 (1947).
stream are used to measure the temperature of the vapor after i t has passed the heater. By measnring the resistances of the thermometers before and after supplying energy to the heater a value of the temperature increase, accurate to a few thousandths of a degree, is obtained. From the mass of vapor flowing through the calorimeter per unit of time, the power supplied to the calorimeter heater, and the temperature increase of the vapor, values of the apparent heat capacities are calculated. The apparent heat capacities differ from true heat capacities because of heat losses in the calorimeter. By making measurements a t four flow rates, coveting the range 0.05 to 0.25 mole per minute, data are obtained by which corrections may be made for heat losses. The design of the calorimeter is such that the relationship C P ( ~ ~= ~C.p) k / F accurately represents the experimental data, where C p is the true heat capacity, F is the flow rate and k a constant for a given set of operating conditions. From this i t follows that a plot of Cp(obs.lvs. 1 / F extrapolated linearly to zero value of 1 / F gives the correct value of CP. To obtain accurate values of Ci, the heat capacity of the vapor in the ideal gas state, heat capacities are determined a t two or more pressures and the results extrapolated linearly to zero pressure. The validity of the linear extrapolation has been checked experimentally for a number of compounds. In the case of 2,2-dimethylbutane the mode of carrying out the heat capacity measurements a t 1 atmosphere pressure differed from the other measurements reported in this paper and from the work reported earlier.2 Instead of maintaining a fixed boiling temperature in the vaporizer, for a series of measurements a t different flow rates, the helium pressure in the system was maintained constant. Because of the pressure head between the vaporizer and the condensers the boiling temperatures were different at different flow rates. This necessitated a correction for the temperature
+
2276
GUYWADDINGTON AND DONALD R. DOUSLIN
coefficient of the heat of vaporization and modified the pressure correction described previously.2 The net result of this difference in experimental procedure is that the consistency between the 1atmosphere measurements and the 293-mm. measurements (carried out one year later) is slightly inferior t o that of other compounds studied with this apparatus. This reveals itself when values of C i - C; are plotted against temperature. Materials--The 2,2-dimethylbutane was Phillips Petroleum Company's "Pure Neohexane." After distillation through a 77-plate column a t a reflux ratio of GO : 1 i t had the following constants : d2040.6490; 1 j b 2 0 ~1.3687. A melting point determination indicated a purity of 99.70 mole per cent. assuming the impurities to be liquid-soluble, solid-insoluble. The n-hexane, while ostensibly Phillips Petroleum Company's Technical Grade, was a special cut3 which, after dearomatization with silica gel and distillation through a 100-plate column a t a reflux ratio of 150 : 1, had the following properties : dZo40.6595, n Z 0 D 1.3748. The purity determined by the freezing point method was 99.77 mole per cent. Heats of Vaporization.-The determination of the flow of vapor is also a method of obtaining heats of vaporization which differs in some aspects from other modern methods capable of yielding accurate results. To check the accuracy of the method of measuring heats of vaporization used in this work, measurements were made on benzene4a t four temperatures, several experiments being carried out a t each temperature. The results of all experiments are given in Table I. Rates of flow were in the range 0.05 to 0.08 mole/ minute, the mass collected varied from 23 to 45 g. and the time of collection was six to eight minutes. It is seen that the values at 49.9'and 80.1' closely check the experimental points of Fiock, Ginnings and Holton6 :it these temperatures. Moreover, it may be shown graphically that the four points are in excellent agreement with a value of 8090 cal./ mole determined by Osborne and Ginnings6a t 25'. A further direct comparison with results of high accuracy is available in the case of 2,2-dimethylbutane. At 25', Osborne and Ginnings6 find, for this compound, a value of 6617 cal./mole. I n the present: work a value of 6640 cal./mole was obtained a t 22.9'. Using a value of - 12.6 cal./ moleldeg. (dcrived from eq. 1) for d(AH)/dT (3) Obtained through the courteous cooperation of J: W. Tooke and A. E. Buell of the Phillips Petroleum Co. (4) T h e purity of the sample as determined from a time-temperature freezing curve was 99.93 mole per cent. Thiophene, a common impurity in benzene, forms solid solutions with benzene hence its presence is not revealed by freezing curve studies. A bomb sulfur analysis did not indicate any significant amount of sulfur. The authors thank M r . E. L. Garton and Miss Shirley Nix of this Station for performing the bomb sulfur analysis. ( 5 ) E. F. Fiock, D. C. Ginnings and W. B. Holton, J. Research Null. Bur. Standards, 6 , 881 (1931). (6) N. S. Osborne and D. C. Ginnings, Null. Bureau of Sfandards, unpublished data.
Vol. 69
TABLE I THEHEATO F VAPORIZATION t, "C.
AHvao.cal./mole
Mean Fiock, et al.5
OF BENZENE?
41.6 7866 7870
49.9 7761 7753 7751 786812 775514 7757
60.9 7606 7604 7608 760612
80.1 7348 7350 7350 7349*1 7353
shows these two results to agree within 3 cal. Fiock, Ginnings and Holton estimate the uncertainty of their measurements to be O.lY0 or less. This would indicate a maximum uncertainty of about 0.15% for the benzene value a t 80.1' from the present work with the possibility of its being less. For other compounds studied the uncertainty in the heats of vaporization should be similar to the benzene results save where impurity in the available sample may have some effect. Table I1 summarizes the experimental values of the heat of vaporization for 2,2-dimethylbutane and n-hexane. Six8 experiments covering a wide range of flow rates are listed for neohexane a t the normal boiling point. The ebullition is extremely violent a t the highest flow rate and could conceivably lead to entrainment of liquid droplets in the vapor stream, with resulting low values for the heat of vaporization. Many investigators have taken precautions against this effect. The evidence here suggests that the presence of an adeTABLEI1 THE HEATSO F
VAPORIZATION OF 2,2-DIMETHYLBUTANE AND %-HEXANE
t,
oc.
49.7
25.0 (ref. 6) 22.9
Flow, moles/min.
AHvsp.,
3 expts.
AHvsp.
calcd. Clapeyrona
6287
6288
6616 6643
6642
cal./mole
2.2-Dimethylbutane 0.057 6283 .088 6294 .112 6292 ,144 6286 ,234 6282 .378 6287 Mean 6287 1 7
.....
AHvap.
calcd. eq. 1, 2
6617 6640
*4
%-Hexane 68.7 4 expts. 6896 1 3 6895 6894 54.8 3expts. 7115 1 3 7115 7116 35.6 3 expts. 7393 * 3 7394 7393 25.0 (ref. 6) ..... 7540 7540 7543 "Since the molal volumes of vapor employed in calculating these values are derived in part from the experimental heats of vaporization of this work and of Osborne and Ginning9 the figures of this column are not an independent check of the experimental values of column 3.
-
(7) In this paper the following values are used: 1 cal. = 4.1833 1.008: int. joules: at. wt. carbon = 12.010; at. wt. hydrogen 0' C. = 273.16' K. (8) The mean valut a n d the mean deviation of the thirteen experiments actually performed is essentially the same as for the six listed
HEATCAPACITIES
Oct., 1947
OF
WHEXANE AND
quate vapor space above the boiling hydrocarbon permits any liquid droplets formed by bursting bubbles to return to the liquid. Kilpatrick and Pitzers have measured the vapor pressure of 2,2-dimethylbutane from 1.9 to 225 mm., while Willingham,10et al., have measured this property from 217 to 780 mm. Each has proposed an Antoine-type equation to represent the data. The consistency of these two equations in the region where they overlap and of the experimentally obtained value of 6640 cal./mole for the heat of vaporization a t 22.9' may be checked by using the Clapeyron equation AHvap. = TAVdP/dT to obtain heats of vaporization from the two vaporpressure equations and from the molal volumes of vapor and liquid. The equation of Kilpatrick and Pitzer leads to 6634 cal./mole while that of Willingham and co-workers yields 6642 cal.fmole. The agreement is satisfactory. Vapor volumes for 2,2-dimethylbutane were calculated from the BP, the value of the secequation PV =; RT ond virial coefficient being taken as B = - 555 6.2 X 103T-l - 7.07 X 10I2 T - 4 ( ~ ~ . )This . is derived from measured heats of vaporization and the variation of heat capacity with pressure as will be shown later in this paper. The use of the Berthelot equation to calculate the volume of the vapor, in the case of this compound, also gives good agreement between calculated and experimental heats of vaporization. The equation of Willingham, et aZ.,l0 in conjunction with the Clapeyron equation and equation of state data referred to above gives a calculated heat of vaporization of 6288 cal./mole a t the normal boiling point in good agreement with the experimental value of 6287 cal./mole given in Table I1 but disagrees with Pitzer's'l experimental value of 6355 cal./mole. The equation AHvsp.= 8149 f 2.37T - 0.0252T2 (1) may be used for interpolation over the range of experimentation. The slight curvature indicated by the equation. was arrived a t by consideration of vapor pressure and gas imperfection data. Table I1 also lists the heat of vaporization of nhexane a t four temperatures, that a t 25' being from the work of Osborne and Ginnings6while the other three are from this research. The values calcd. are derived from the Clapeyunder AHvap, ron equation, using the vapor pressure data of Willingham, et al.,10and the vapor volumes are obtained by use of a second virial coefficient
+
+
2277
2,%DlMETHYLBUTANE
a heat of vaporization 0.75% higher than the experimental value at the normal boiling point. The empirical equation AHvsp.= 8571
+ 6.372T - 0.03298T'
(2)
is in almost perfect agreement with the experimental values for n-hexane from 25' to the normal boiling point but is not of a form to give accurate values outside of the range of experiment. The figures in column 4 of Table I1 are from equations (1) and (2). Lemons with Felsing12 have measured the heat of vaporization of n-hexane over the temperature range covered here with an estimated accuracy of 1%. Their values average about 1% lower than those of this paper. Vapor Heat Capacities.-Table I11 gives the experimental values of the molal vapor heat capacities of 2,2-dimethylbutane and n-hexane a t the various temperatures and pressures used. A t most temperatures, measurements were made at two pressures so that C;, the heat capacity of the gas in the ideal gas state, could be obtained by linear extrapolation to zero pressure. For n-hexane a t 365.15' K., measurements were made a t three pressures and show that, in the pressure range studied, the heat capacity of the vapor is a linear function of pressure. This fact also has been demonstrated for n-heptane.2 TABLEI11 EXPERIMENTAL HEATCAPACITIES, CAL./DEG./MOLE 2,2-Dimethylbutane P.. mm.
293.4 760.0
T,O K .
+
341.55
353.20
376.05
412.40
449.40
38.400 41.665 45.075 48.410 38.870 39.880 41.930 45.275 48.535 n-Hexane
p.,
mm.
235.7 479.4 760.0
T ,OK. + 333.85
365.15
398.85
433.70
468.90
37.740 40.440 43.430 46.460 49.505 38.140 40.685 40.950 43.735 46.630 49.600
Since the results are highly precise and since the evaluation of C$ - C: depends on differences, values are given to 0.005 cal./mole/deg. although the absolute accuracy does not approach this. As with other compounds studiedI2 large scale plots of C p vs. T for the 1 atm. results for the two compounds give smooth curves slightly concave toward the T axis a t high temperatures but tending to curve in the opposite direction a t low temB = -2270 .+ 6.25 X 10'T-I - 1.306 X 1 0 I 3 T 4 (cc.), peratures due to the greater contributions of gas imperfection in this region. The close agreement implies excellent consistency Heat Capacity in the Ideal Gas State.-From of the data from three laboratories. the data of Table 111, values of C i were obtained Use of the Berthelot equation for calculating the molal volume of the vapor of n-hexane leads to at each temperature by linear extrapolation from the experimental points to zero pressure. These (9) J. E. Kilpatrick and K. S . Pitzer, THIS JOURNAL, 68, 1060 are listed in Table IV along with values of Ci, (1946). (10) C. B. Willingham, W. J. Taylor, J. M. Pignocco and F. D. the heat capacity a t 1 atmosphere. The following Rossini, J . Rcscarclt N o f l . Bur. Standards, 36, 219 (1945). 68, 2413 (1941). 111) K . S . Pitzer, THISJOURNAL,
(12) J. F. Lemons with W. A. Felsing, ibid. 66, 46 (1943).
GUYWADDINGTON AND DONALD R. DOUSLIN
2278
Vol. GO
Gas Imperfections.-Two aspects of the foregoing experimental work are intimately related to gas imperfections. The molal volume of the vapor is related to the heat of vaporization by the exact Clapeyron equation A H v a p . = T( V, - V I ) d P / d T while the variation of heat caDacitv of n-Hexane : the vapor with pressure is given by (bCpjbPfT = - T(b2V B T ? P . c; = 4.:180 + 0 . 1 0 5 7 3 ~- i.9948 x io-T* (3) From ’the ikeasured values of AHvap.and of 2,S-Dimethylbutane: ( A C ~ / A P ) Ti t is possible to determine the temC; = -2,558 + 0.13724T - 5.339 X 1OW6T2(4) perature dependency of B, in the equation of state P V = R T BP. For a low-pressure vapor, TABLEIV HEATCAPACITIES IN THE IDEAL GASSTATE,CAL./DEG./ B can be identified with the second virial coefficient with little error.I6 The dependence of heat MOLE capacity on pressure in terms of the above equac; - --c; . -T(d2B/ tion of state becoinrs ( b C p / b P ) ~= T , “K. Ck C> ICxpt. 2nd Virial Berthelot b1‘2)p. %-Hexane If B is set equal to Bo - Ao/l’ - Co/’l’” the 333.85 38.60 37.35 (1.25) (1.25) (0.59) variation of heat capacity with pressure a t con365.15 40.95 40.22 0.73 0.75 .45 stant temperature is 2Ao/T2 n(n 1)Co/Tn+’. .44 398.85 43.74 43.30 .44 e35 The value of n which gives a linear plot of T2.24 .25 .27 433.70 46.63 46.39 ( ACp/ AP)T vs. l/Tn-’ will give a good fit of the 468.90 49.80 49.46 .14 .14 .21 experimental data. For 2,2-dimethylbutane and n-hexane n equals four. From this same plot A0 2,2-Dirnethylbutane and Co may be readily evaluated. The values of 341.55 38.87 38.10 0.77 0.73 0.50 (ACp/ AP)r are numerically equal to Ck - C i 353.20 39.88 39.25 * 45 since i t has been demonstrated experimentally 376.05 41.93 41.50 .43 .45 .37 that Cp is a linear function of pressure a t constant 412.40 45.28 44.95 .33 .29 .28 temperature. To obtain the remaining constant, 449.40 48.64 48.33 .21 .19 .22 Bo, in the expression for B , a molal volume of the Pitzer“ has determined the heat capacity of vapor may be determined from a heat of vapori2,2-dimethylbutane vapor a t three temperatures. zation measurement] by use of the exact ClapeyThe scattering of his points from the results of this ron relationship, which permits B to be evaluated work is of the magnitude of his estimated 1% ac- from the equation of state. BOcan now be found curacy. Eucken and Sarstedt’s’3 point a t 451’ K. by difference. In this way, using all the measured is about 0.476 higher than the value reported here heats of vaporization of Table IV to obtain mean which is well within their claimed 27, accuracy. values of Bo, the following equations of state were The accuracy of this work is estimated as prob- determined. ably within *0.2%, n-Hexane: P V = R T - (2270 - 6.25 X lo6 T-’ + 1.306 In a recen.t paper Kilpatrick and Pitzerg comx 1013 ~ - 4 p (5) pare C: values derived from the one atmosphere 2,2-Dimethylbutane: P V = RT - (567 - 6.2 X 103T-’ f7.07 X 1OI2 Y 4 ) P (6) measurements reported here with their statistically calculated heat capacities of 2,2-dimethylbu- The units are P in atm., V in cc. and T in O K. tane. The correction for gas imperfection was In Table IV the results of calculating C$ made by use of the Berthelot equation of state. C$ from the above equations are given in the colMeasurements made later a t 293.4 mm. showed the Berthelot correction to be inadequate, the dis- umn headed “2nd Virial.” The excellent agreecrepancy amounting to about 0.25 cal./deg./mole ment with experimental values for n-hexane is an a t the lowest temperature. This change brings indication of the precision of the measured values the slopes of the calculated and experimental of CP. It is also seen that use of the Berthelot equation does not give highly accurate values of values into closer agreement. Eucken and Sarstedt’s13value of C) for n-hex- C$ a t low temperatures. In using the Berthelot equation Kay’s’’ values ane a t 451’ K . is 2% higher and Bennewitz and Rossner’s14value a t 410’ K. is six calories lower -for the critical constants of n-hexane and 2,B-dithan those of Table IV. Pitzer’s’6 values calcu- methylbutane were used. The calculated values lated by approximate statistical methods range of Cb - C: for 2,2-dimethylbutane indicate from one cal./deg./mole higher a t the lower end somewhat less precision in the measured values of of the temperature range to about 0.3 cal.ldeg.1 Ck - Cg for this compound. This is probably bemole higher a t the upper end of the range. cause the l-atmosphere measurements were made equations may be used to obtain values of C: by interpolation without loss of accuracy, since the maximum deviation of an experimental point from the curves represented by the equations is 0.03570.
+
-
+
(13) Eucken arid Sarstedt, Z . physik. Chem., BSO, 143 (1941). (14) K. Bennewitz and W. Rossner, ibid., B39, 125 (1938). (15) Kenneth El. Pitzer, Ind. E n g . Chem., 36, 829 (1944).
+
(16) Charles F. Curtis and J. 0. Hirschfelder, J . Chcm. P h y s . , 8 , 491 (1942). (17) Webster B. Kay, THISJOURNAL, 68, 1336 (194fi),
Oct., 1947
2279
THERMAL DECOMPOSITION OF %-HEPTANE
a year earlier than the results a t lower pressure, using the somewhat less satisfactory experimental procedure referred t o previously. It is evident that equations ( 3 ) and (4) in conjunction with equations ( 5 ) and (6) may be used to obtain the heat capacity of either compound at any temperature and pressure in the range studied. Probably extrapolation to pressures of several atmospheres would give results adequate for many purposes. The calculated heats of vaporization in Table I1 provide a further excellent check of the consistency of equations ( 5 ) and (6). The calculated values of the heats of vaporization of 2,2-dimethylbutane were obtained by use of a slightly modified version of eq. (6) derived from Osborne and Ginnings6 25’ heat of vaporization and the variation of C i with temperature, thus providing calculated values of the heats of vaporizatibn a t 22.9 and 49.7’ which are essentially independent of the heats of vaporization measured in this Laboratory. Acknowledgments.-We wish t o thank hlr. R. M. Gooding of this Station for the time and care spent in the purification of the compounds. We also appreciate the assistance given by hlr. Harold Coleman to Mr. Gooding in the determination of purity and physical constants. The valuable suggestions made by Dr. S . S. Todd during the progress of this work and his assist-
ance in making some of the measurements are gratefully acknowledged. The authors appreciate the encouragement and helpful advice given by Dr. H . M. Huffman. Summary The heat capacities of the vapors of n-hexane and 2,2-dimethylbutane have been measured a t several pressures in the temperature range 33 to 470’ K. *he experimental values of the heat capacities of the vapors in the ideal gas state deviate from the following empirical equations by less than 0.1%
+ 0.10573T - 1.9948 X
n-Hexane: C$ = 4.280 2,2-Diinethylbutane: C) = -2.558
10-5T2
+ 0.13724T - 5.339 X lO-6.T”
The heats of vaporization of the two compounds have been measured and are represented with an accuracy of approximately *O.l% in the range from 25’ to the normal boiling points by the equations
+
-
n-Hexane: = 8571 6.372T 0.03298T2 2,2-Dimethylbutane: AHvap.= 8149 2.37T = 0.0252P
+
Second virial coefficients which represent the low temperature, low pressure behavior of the vapors have been obtained from the thermal measurements. BARTLESVILLE, OKLA.
RECEIVED APRIL 7, 1947
__
_______.
[ COKTRIRUTION FROM
TIIE
RESEARCH LABORATORIES, HOUSTON REFINERY,SHELL OIL
COMPANY, INCORPORATED]
Kinetics and Mechanism of the Thermal Decomposition of n-Heptane1 BY W. G. APPLEBY,W. H. A V E R YAND ~ ~ W. K. MEERBOTT Introduction During the past decade a large number of papers has appeared in the literature which deal with the thermal decomposition of paraffin hydrocarbons. Because of the labor involved in the analysis of the decomposition products, very few investigations have been reported in which coniplete enough analytical data were obtained to permit conclusions of general validity to be drawn regarding the kinetics and mechanism of the decomposition of the normally liquid paraffins. If the Rice free radical chain theory predicts the products formed in the decomposition of paraffins, then no saturated compounds other than hydrogen, methane, ethane and propane should be iornied in the pyrolysis a t atmospheric pressure of parafEn hydrocarbons. This, in turn, would imply that the amount of unreacted paraffin remaining after pyrolysis of a liquid paraffin could be (1) T h e work described here was carried out during 1942. Publication has been de‘!ayed by the pressure of other work during t h e war. ( l a ) Prestmt nridi.ew: Johns Hopkins Vniversity. Silver Spring, hfcl.
determined from the weight of the liquid product remaining after suitable treatment for removing olefins. The hypothesis that, after treatment of the liquid sample to remove olefins, only unchanged charging stock should remain can be supported by determination of the refractive index of the treated sample. This line of reasoning has been tested by Dintses and co-workers.2 Pure normal octane was pyrolyzed to 20y0 decomposition in a copper reactor. The liquid and gaseous product showed on “most careful analysis” no saturated hydrocarbons other than methane, ethane and propane. Further tests were made on fractions of Baku gasoline (containing 60% naphthenes) that had been treated with concentrated sulfuric acid before pyrolysis. In every case, treatment of the liquid product of the reaction with sulfuric acid left a material identical in refractive index and other physical properties with the original charging stock. (2) A. I . Dintses and A. V. Frost, J . Gcn. Chem., U. S.S. R., 8 , 747 (1933), 4, G1O (1934); A I. Dintses and A. V. Zherko, i b i d . , 6, 68 (19363.
